Hydrogen Paste Meets Reality: Energy In, Energy Out, And What’s Missing

Someone suggested Fraunhofer’s Powerpaste as a solution to the safety and cost challenges of hydrogen at sea after I wrote about the new DNV maritime hydrogen safety guidelines that make it even more uneconomic. It is an understandable instinct. If compressed hydrogen is dangerous and expensive, perhaps changing its form solves the problem? The question is whether this is a different path or simply the same physics rearranged. My assumption was the latter or worse, having looked at almost all of the dead ends for economically viable hydrogen storage over the past decade. I wasn’t surprised to find that it was in the worse category, having very high costs and no possible use cases that couldn’t be met much more cheaply and simply by other technologies.

Powerpaste was developed by Fraunhofer IFAM over the past couple of decades as a hydrogen carrier based on magnesium hydride, and unveiled in 2021. The concept is simple on the surface. A paste containing magnesium hydride reacts with water to produce hydrogen gas and magnesium hydroxide. The hydrogen feeds a fuel cell and produces electricity. The system avoids high-pressure tanks and operates at ambient conditions. It is presented as a cartridge-based solution that is easy to handle and transport.

It sounds promising because it appears to solve hydrogen’s hardest problem. Compressed hydrogen requires 350 to 700 bar storage. Liquid hydrogen requires cryogenic temperatures. Powerpaste avoids both. It also claims higher volumetric energy density than compressed hydrogen. It can be stored and transported more easily. It appears to offer a practical way to deploy hydrogen in small systems.

The first constraint appears when you look at what is actually being carried. Fraunhofer states that about 10 kg of paste produces 1 kg of hydrogen. The reaction requires water, and the stoichiometric requirement is about 9 kg of water per kg of hydrogen. That means that producing 1 kg of hydrogen requires roughly 19 kg of input materials. About half of the hydrogen comes from the water, not the paste. It can’t be just any water, and more on that later. This is not a storage medium in the conventional sense. It is a paired chemical system where both reactants must be transported.

The commonly cited energy density numbers reflect only the paste. Values of about 2 kWh per liter and 1.6 to 2 kWh per kg are quoted. Once water is included, those numbers drop sharply. Using Fraunhofer’s own framing of about 16 kWh of electricity per kg of hydrogen, the combined paste and water mass of about 19 kg yields about 0.84 kWh per kg. The combined volume is similar, giving about 0.8 to 0.9 kWh per liter. But that is still not the system a user actually carries. A complete unit must also include the fuel cell, reactor, pumps, filtration, hydrogen conditioning, thermal management, power electronics, and cartridge structure. When those are included, a realistic system lands closer to about 0.3 to 0.4 kWh per kg and about 0.25 to 0.4 kWh per liter. That places it in the same range as complete modern battery systems rather than above them, and represents an order of magnitude overstatement in Fraunhofer’s claims. Gasoline is about 9 to 10 kWh per liter. Compressed hydrogen systems are about 1.3 kWh per liter. The headline advantage shrinks by 90% once the full system is counted.

The next constraint is hidden in the magnesium. The paste is based on magnesium hydride, and only half of the hydrogen comes from water. The other half comes from the hydride. Producing 1 kg of hydrogen requires about 6 kg of magnesium metal embedded in the paste. Magnesium production is energy intensive. Electrolytic processes require about 14 to 18 kWh per kg of magnesium. Multiplying that by 6 kg yields about 80 to 110 kWh of electricity just to produce the magnesium required for 1 kg of hydrogen output. Some industrial processes are higher. This energy is not visible in the product, but it is real and must be paid. The destruction of magnesium in this process means that the energy to make it must be included in the energy balances.

The full system efficiency reflects all of these steps. Electricity is used to produce hydrogen through electrolysis at about 65% to 75% efficiency. Hydrogen is used to form magnesium hydride. The paste is transported and then reacts with water to release hydrogen. The hydrogen feeds a fuel cell at about 45% to 55% efficiency. Even without including magnesium production, the electricity to electricity efficiency is about 30%. Including magnesium production, the effective energy return is in the 10% range because over one hundred kWh have already been consumed upstream. This is not a storage loop. It is a multi-step conversion chain with significant losses at each stage.

The reaction itself introduces another constraint. Magnesium hydride hydrolysis releases about 19 kWh of heat per kg of hydrogen released. This heat must be managed. A system producing 1 kg of hydrogen over 10 hours would need to dissipate about 1.9 kW of thermal energy. Faster production rates increase the thermal load proportionally. This makes the system behave like a chemical reactor rather than a passive storage system. It requires thermal management, controlled dosing, and pressure regulation. This means that getting that single kg of hydrogen out takes hours, not minutes or seconds. That means that the feasible power that can be delivered from Powerpaste is measured in watts or per single kilowatts.

The hydrogen produced is not immediately suitable for a fuel cell. The reaction takes place in contact with water and produces a wet gas stream. It can contain water vapor, droplets, and particulates from the reaction. PEM fuel cells require hydrogen purity above 99.97% with strict limits on contaminants. Even small amounts of impurities can degrade performance and shorten lifespan. This means the system requires gas-liquid separation, drying, and filtration. These components add cost, weight, and complexity. And back to the water: it requires virtually pure water. Tap water, brackish water, or contaminated water could all raise the risk of aerosols, dissolved salts, and trace contaminants unless the system includes very good gas-liquid separation and purification stages. This eliminates the Fraunhofer claim of any water being suitable and means that water must be counted in the energy density calculations. Pure water must be carried with the solution, or an entire high end water purification system must be added to the stack, or the fuel cell will fail rapidly.

The cost structure reflects both visible and hidden elements. Fraunhofer suggests about €2 per kg of paste, which translates to about €20 per kg of hydrogen output. This excludes water handling, system hardware, purification, and logistics. It also excludes the cost of magnesium production energy. If 80 to 110 kWh of electricity is required to produce the magnesium, and electricity costs $0.05 to $0.10 per kWh, that alone adds $4 to $11 per kg of hydrogen before any other costs. The full system cost is higher.

The recycling loop is not resolved. The reaction produces magnesium hydroxide. Converting this back to magnesium metal requires high-temperature processing and significant energy, pretty much the same as refining the magnesium in the first place. There is no widely demonstrated closed-loop system with competitive economics. In practice, this means either consuming magnesium as a material input or relying on energy-intensive recycling that doesn’t exist. This is very similar to the fundamental problem of aluminum air batteries, where effectively you destroy aluminum to release energy, then have to send the result back to the aluminum foundry.

Scaling this to maritime applications shows the mismatch clearly. A small ferry can require tens of MWh of energy per trip. At 0.84 kWh per kg including water, delivering 10 MWh would require about 12,000 kg of paste and water combined. The system would also need to manage about 12 MWh of thermal output during operation. This is not a storage system. It is a large-scale chemical logistics and thermal management problem.

Even in small applications, the system struggles. Fraunhofer has demonstrated prototype scooters and portable power units in the 100 W to 1 kW range. These are theoretically logical targets because compressed hydrogen is impractical at that scale. However, these remain demonstration systems. The hydrogen purity requirement means that even small units need gas conditioning equipment. The thermal and reaction control requirements add further complexity. The result is a system that is much more complex and expensive than a battery and solar panels for the same energy output. Remember, there is no recharging of the Powerpaste cartridges, they all have to be delivered to the site. Meanwhile, throw some dirt cheap solar panels up beside a battery and get a lot more power and a lot more energy for a lot less cost. And if you want to power a scooter, do the same thing, put batteries in it.

Battery systems provide a useful comparison. Modern battery packs deliver 0.2 to 0.3 kWh per kg and similar volumetric density. They do this without chemical conversion, without gas handling, and with round-trip efficiencies of 85% to 95%. They are simple to operate and are improving in cost and performance. They scale to GW capacity, or can be delivered in tiny cells. Powerpaste offers similar energy density with lower efficiency and greater complexity. The advantage is illusory.

Powerpaste is best understood as a hydrogen carrier optimized for handling convenience rather than system performance or cost. It replaces high-pressure storage with chemical storage. It avoids some challenges but introduces others. It shifts energy consumption upstream into material production. It adds conversion steps that reduce efficiency. It just moves around inside the failure space, not outside of it, as with all hydrogen storage solutions. In this quadrant view of hydrogen storage approaches I developed for the life story of a committed hydrogen for energy researcher, it’s just another expensive option with even less utility.

Hydrogen advocates bring it up because it appears to address an intuitive concern. Hydrogen storage is difficult, so a solid or paste form appears attractive. The headline metrics focus on the paste itself and exclude water and upstream energy. The incredibly low power output is ignored. This creates an impression of higher performance than the full system delivers. The appeal is understandable and it’s obvious why people suffering from confirmation bias thing it is a great solution, but the numbers do not support it. Anyone suggesting it as an energy carrier for ferries clearly has applied no critical thinking, analytical skills or basic research to it.

Fraunhofer’s presentation of Powerpaste relies on a set of claims that are technically defensible only under narrow, incomplete boundaries, but which are misleading when interpreted the way most readers will interpret them. Statements about energy density, system performance, simplicity, and cost exclude the required water, exclude the upstream energy for magnesium production, and exclude the gas conditioning and balance of plant required to make the hydrogen usable in a fuel cell.

Taken together, these omissions create an impression that Powerpaste outperforms batteries and offers a practical hydrogen solution, when a full accounting shows at best comparable energy density, far lower capacity, far lower energy storage,  far lower efficiency, and materially higher complexity and cost. The result is that hydrogen advocates and non-specialists alike are left with a deeply distorted understanding of the technology’s capabilities. Fraunhofer is a respected applied research organization, and it should present Powerpaste with the same system-level rigor it applies elsewhere, including full input accounting and realistic performance boundaries, rather than continuing to promote it in a way that overstates its relevance and obscures its limitations.

When evaluated end to end, Powerpaste is an interesting piece of chemistry that does not translate into a compelling energy solution. It cannot match batteries on cost, real energy density, efficiency or simplicity. It cannot not scale beyond tiny systems. It embeds significant upstream energy costs. It remains a demonstration of what is technically possible rather than a solution that meets the requirements of real-world energy systems. It’s energy destruction, not energy storage. It’s unclear why Fraunhofer continues to work on it despite its insurmountable thermodynamic failings. It is clear why hydrogen for energy types would point to it in desperation, however.